For a large number of systems and estimates I would propose the following approach.
Categorize the systems so that optimization can be run for up to 100 systems in each category and then over the category index values in order to overcome the 100 limit for the optimization input.
Ascertain the available historical cash flow data for each system in order to determine the appropriate frequency (months, quarters etc.) for calculation of historical volatility.
Use the option to modify the correlation matrix for the optimization and then update the expected return data to reflect the estimated budgets that are available for each system.
Combine all individual category optimizations by running optimization on category groups and multiplying each system by the capital allocated to the parent category.
When a system comes up for review, optimization can be evaluated by comparing the category (or entire organization) optimization results with the optimization run without that system. This will evaluate the value added of that system to the category or organization which can then be used for decisions on whether or not to continue with the system.
